Disable hipcc optimizations i.e. use -O2 instead of -O3 (work around for gq_ttq crash 806 on AMD GPUs at LUMI)#1007

Merged

valassi merged 14 commits intomadgraph5:masterfrom

Sep 25, 2024

Commits on Sep 19, 2024

[amd] in gq_ttq.mad HiprandRandomNumberKernel.cc, add debug printouts (commented out) for the memory corruption #806
valassi
committed
[amd] in gq_ttq.mad MemoryBuffers.h, temporarely use c++ malloc instead of HIP pinned host malloc to debug #806 - still crashes, will revert
valassi
committed
[amd] in gq_ttq.mad timermap.h, add some debug printouts for the crash #806 - now valgrind gives no invalid read, but there is a 'Memory access fault'
valassi
committed
[amd] in gq_ttq.mad cudacpp.mk, enable -ggdb... the issue seems to be in vxxxxx (which may explain why this only appears in gqttq?)
valassi
committed
[amd] in gq_ttq.mad, revert the last four temporary changes introduced for debugging the crash #806 in hipcc
valassi
committed
[amd] in gq_ttq.mad and CODEGEN, work around the memory access crash #806 for HIPCC by disabling hipcc optimizations (use -O0 instead of -O3)
valassi
committed
[amd] in gq_ttq.mad and CODEGEN, work around the memory access crash #806 for HIPCC by disabling hipcc -O3, but keep -O2 (better than -O0)
valassi
committed
[amd] regenerate all processes, including OPTFLAGS=-O2 for hipcc instead of -O3 (workaround for gq_ttq crash #806)
valassi
committed

Commits on Sep 20, 2024

Commits on Sep 21, 2024

[amd] move back to previous upstream/master codegen logs to ease future merges
valassi
committed